TWI375214B - System for simulating human singing and method thereof - Google Patents

Description

IX. Description of the invention: [Technical field to which the invention pertains] A speech simulation system and method thereof, in particular, a system for simulating vocal singing and a method thereof. [Prior Art] Singing has become one of the rest activities of modern people. Now, in addition to singing to KTV using sing-along equipment, there are also home sing-along devices that provide singers to sing at home. Whether singing at κτν or using a vocal device at home, the singer will see the lyrics of the song on the screen or on the TV, and will hear the tune of the song, namely the song _ rhythm and melody, so that the singer can follow the tune The lyrics sung by the original singer of the songs displayed on the screen or on the TV. In some cases, the singer does not know how the song sings the singer and will sing the voice of the vocalist. The sound of the original singing song can be accompanied by the song's tune—touching the silk, letting the singer sing together, so 'the songs stored by the singer device currently used in κτν^, in addition to storing the lyrics of the song, the tune of the song. In addition to the song record, the singer and the singer are also singed, and the original singer’s song is also stored. The song of the original singer and the song of the song are played by the _ channel. The song will play the original singing voice, ± 静音 will mute the channel of the original singing voice, so the singer will not hear the original singer's voice. However, because the number of songs is too large, the original singer's singer is bound to be "the rich tfr song-like sing-along equipment provided by 财大_ _ empty space is not big enough, two, and '- the song of the original singer of all songs ^ The function of the guide cannot be saved. , ~, earth ... method for all songs to provide a vocal summary, we can see that the prior art has long-term all songs to provide vocal vocal function J ...., the technical means to solve this - problem / therefore it is necessary to propose improvements f Inventive content] There is no singularity in the prior art for the vocal singing of all songs:::: The invention discloses a system for simulating vocal singing and the system for simulating vocal singing disclosed by the present invention, the application In the accompaniment device, the vocal device outputs the lyrics of the song and the tune of the song, which includes: a storage module for storing a plurality of voice files, each voice file corresponding to the text; the input module for rotating the song The lyrics and the tunes of the songs; the voices rely on the group, the ribs are extracted from the voice files corresponding to the cultural relics of the songs, respectively, corresponding to the voice signals of the respective texts; the voice adjustment module 'the tone of the voice signals for the ages The long-term (four) and the corresponding texts have the same pitch and play length in the tune; the speech synthesis module' is used to combine the voice signals according to the order of the characters in the lyrics. The voice-being 'voice output module' is used to turn out the voice data. It uses the voice 撷1375214 模财巾(4), and adjusts the tone and length of the voice signal after the voice adjustment. :,: Select the order of the characters in the lyrics to synthesize the adjusted °. 9 News turned to the problem of the prior art. The method of simulating vocal singing by the invention is applied to the vocal chorus. The chorus device outputs the lyrics of the song and the tune of the song, including the voice storage of the text of the chest.

The lyrics of the song and the tune of the song; the voices corresponding to the characters of the songs correspond to the voice signals of each character; the pitch and the length of each voice are adjusted to correspond to the corresponding characters in the tune. The corresponding pitch and play length are the same; according to the order of the text in the lyrics, & is corresponding to the „.g condition number is the voice data; the output voice data is obtained by capturing the voice signal of each text of the corresponding song, and according to After the tone of the whole voice signal and the length of the play, according to the arrangement of the characters in the lyrics

The sequential synthesis of the adjusted speech 峨 is based on the problems of the prior art. The system and method disclosed in the present invention are as above, and the difference between the prior art and the prior art is that the present invention first captures the voice signal corresponding to each character in the lyrics of the song, and plays the tone according to the sound of the song. After the length, 'by the lion: the various characters of the text are synthesized in sequence to synthesize each adjusted voice. Through the above technical means, the present invention can provide all the songs that the singer will sing, and achieve the technical effect of the vocal vocal singing. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The detailed features and embodiments of the present invention will be described in detail below with reference to the drawings and embodiments, which are sufficient to enable those skilled in the art to readily understand the technical means to which the present invention solves the technical problems. And understanding the achievable effects of the present invention. The following is an illustration of the environment for providing simulated vocal singing in "Picture 1". The present invention is implemented in a accompaniment device 100, and the accompaniment device 1 影像, the video playback device 600, and the sound playback device 700 are provided to provide a singer for the user to sing a song. The singer device 1 includes but is not limited to a point. A player, a computer that performs a singing service on a webpage, and a video playback device include, but are not limited to, a television or a computer screen. The accompaniment device 1 is connected to the video playback device 600 and the audio playback device 7A. In the following, the system operation of the system for simulating vocal singing proposed in the "Fig. 2" of the present invention is first described. As shown, the system of the present invention includes a storage module 11A, an input module 12G, a voice operation module 130, a voice adjustment module 150, a voice synthesis module 17A, and a voice output module 180. The storage module 110 is responsible for storing a plurality of voice files, each stored voice file is at least one text, and each text can correspond to multiple = same voice files. The text mentioned in the present invention includes but not a word in a language such as a Chinese language or a Chinese language; the voice slot case proposed by the present invention includes a text, a solution word, or a her sentence, that is, a voice. The pronunciation of one or more words can be recorded in the building case. I S3 8 1375214 General and &, s, my tone building case is based on the word sound and text to establish a corresponding relationship, that is, regardless of whether the different text fonts are the same, as long as the word sounds the same text will correspond to the same voice file However, the present invention is not limited thereto, and a one-to-many correspondence relationship may be established for characters and voice files having different font shapes. In addition, because the same text in different positions in a sentence and the sentence needs to be different from the strong focus 'usually there will be no-like tone, which will make the pronunciation slightly different, for example, "I" is a commonly used text. "The words of "I" are used by ordinary people and "I", especially in the beginning of the speech." The words "I" in the three sentences are different in the characteristics of the voices. There will be at least three pronunciations, so the storage module 110 will store the three voice files corresponding to the text "I".

The input module 120 is responsible for inputting the lyrics of a song and the tune of the song. The singer can sing along with the tune and lyrics of the song being played, so the lyrics have a specific correspondence with the tune, that is, the song === text will correspond to the tune of the tune, and the tune corresponding to each character Contains more than one pitch. In addition, each character rides the voice tones and is recorded in the lyrics, but the lyrics record the texts of the lions, and the pots are limited to the recording start time and the end time ^ /, his way, for example, only record each character volley The playback time is adjusted by the cumulative calculation, and the playback time of the tone corresponding to the text of the different banknotes is I S1 9 to obtain the playback start time of the tone corresponding to a certain character. In I, this is the way to limit and record (4) her, but the invention is not in addition to the upper view of the dragon's voice, the lyrics can be divided into multiple paragraphs, each paragraph contains one or More than the text ^ so, recorded in the time of the song _ for the length of each paragraph _ put. In most cases, the lyrics and tunes will be stored in the storage flip 110 in the two input mode, the group m is selected from the storage module 110 according to the singer's songs, and the present invention does not This is limited to, for example, input 槪12〇 can also provide miscellaneous lyrics and _. The voice release amount 130 is responsible for reading each character in the lyrics after inputting the lyrics in the input module m, and to the storage module 1 i 0 by the voices corresponding to the respective characters in the lyrics (4) Each voice signal taken out by the voice Namo and New 130 corresponds to each text. If the voice file records the pronunciation of the text, the voice signal recorded in the voice file stored in the storage module 110 is the voice signal corresponding to the text, so the voice capture module 130 can complete the complete voice. 5 The voice file is used as the 5 voice signal, and if the voice file records the pronunciation of a word or a short sentence, the voice capture module 130 extracts part of the voice building as the sound sfl number, and is captured. The voice signal that comes out is the pronunciation of the corresponding text. In general, the voice capture module 13 撷 extracts the voice signals corresponding to the respective characters from the voice files corresponding to the respective characters according to the order of the characters in the lyrics, but the voice capture module 130 captures the corresponding lyrics. The order of the voice signals of the text is not limited thereto, and any manner in which the voice capture module 13 can extract the voice signals corresponding to all the characters in the lyrics can be used by the present invention. If the voice file and the text stored in the storage module 110 have a many-to-one correspondence, the voice capture module 130 may use a Hidden Markov Model (HMM) to select a voice corresponding to the text. Signal, but the way in which the present invention selects the voice file to be retrieved is not limited thereto. If the voice file and the text stored in the storage module are in a one-to-one correspondence, the voice file and the text recorded in the storage module 11 are used in the storage module 11〇, _ The voice signal corresponding to the text is retrieved from the voice file corresponding to the text. It is worth mentioning that if the speech capture module 13 uses the hidden Markov model to select the audio file, it is necessary to provide the training data to train the hidden Markov model, and let the hidden Markov model calculate the MFCC. The hidden Markov model can train the different situations of the model from the training data' and each state in the pin-type type trains the conversion function, and then inputs the text and the corresponding voice file as input parameters. After the Markov model, the 'hidden Markov model' will be selected to provide the voice adjustment module 150 with a light-tone blue voice. In addition, the input voice file is a word or phrase that contains words, and the hidden Markov model uses the Viterbi algorithm to obtain the best of the 1375214 state sequence of the input speech age. The sequence allows the hidden Markov model to cut the pronunciation of the voice file. That is to say, the voice capture module 130 can extract the voice signal of the corresponding text from the voice file of the recorded text through the hidden Markov model. The voice adjustment module 150 is configured to adjust the tone of the voice signal extracted by the voice capture module 13 to the same tone as the corresponding text in the tune, and then the voice adjustment module 15 will also transmit the voice signal. The playback length is adjusted to be the same as the playback length of the corresponding tone in the tune. Generally, the voice adjustment module 150 adjusts the pitch of the voice signal according to the sequence in which the voice capture module 130 extracts the voice signal, but the present invention is not limited thereto. If the voice capture module 1; 3 is taken out - the tone of the audio signal is the same as the corresponding tone in the tune before being adjusted by the voice adjustment module 150, then the voice capture is performed. The module 130 will not adjust the voice signal extracted by the voice capture module 13. In general, the voice adjustment module 15 will use the Time Domain Pitch Synchronous Overlap-add (TD-PSOLA) algorithm to adjust the pitch of the voice signal. The time domain fundamental frequency synchronization overlay algorithm will first Calculate the position of each peak in the voice signal 're-adjust the pitch of each peak to complete the adjustment of the pitch of the voice signal'. However, the manner of adjusting the pitch of the voice signal proposed by the present invention is not limited thereto. Other methods such as a phase vocoder and the like. IS] 12 The speech synthesis module 170 is responsible for adjusting the voice adjustment module 15〇, the voice of the tone is synthesized into a voice = Lai, and the towel and the voice group are 17 according to the order of the characters in the lyrics. Corresponding voice signal synthesized speech. If the voice touch 13G extracts the voice signal from the corresponding voice Tan case according to the order of the characters of the song, the voice adjustment module 15G is also extracted according to the voice signal, that is, the text in the lyrics. The voice synthesizing module 170 can directly synthesize the voice signal according to the sequence of the voice adjustment module 15_the whole voice signal; if the voice operation module 13〇/the voice adjustment module is If you do not follow the order of the words in the lyrics, you can manipulate/adjust the voice money. Then the voice depends on 1Γ7_σ-you need to read the input--person model surplus_- 120 input lyrics, so as to obtain the order of the characters in the lyrics, so that The adjusted voice signals of the voice adjustment module 15 can be synthesized in the correct order. It should be specially noted that the length of the playback of the voice signal adjusted by the voice adjustment module 15 in the voice data and the tone corresponding to the voice corresponding to each voice signal are played in the tune. the same. 13 The speech data generated by the 'speech synthesis module 17' will sing the song with the _ 吟, but the generated speech data does not contain the tune, only the vocals. The voice output module 180 is responsible for rotating the voice synthesis module 17 to synthesize the voice data generated by each of the five voices §fl. The voice output module 18 can play voice data, store voice data as a new file, and output voice.

Information, but the invention is not limited thereto. If the voice output module 180 outputs the voice data in a play mode, the accompaniment device 100 executing the present invention directly outputs the voice data and the tune of the song to the connected sound playback device 700, and if the voice output module 18输出 Output voice data in the form of a file. The singer device will output the voice data and the tune of the song to the connected sound playback device 700 after reading the output file. In addition, while the vocal device 8 (8) outputs the voice material and the tune of the song to the connected sound playing device 7 ,, the vocal device 1GG also plays the song _ song feeding (four) f 彡 like Wei device _ play 'making the song The songs, tunes, and synthesized voices can be played simultaneously. 'Providing--using the vocal vocal-function-energy...................

In fact, the input module 12 of the present invention is more responsible for providing the user with the input of the if parameter. The If parameter is provided to the effect module 160 that can be attached to the present invention as a basis for adjusting the input effect of the voice signal. Under the emotion expressed by the non-emotional parameter, the sound and the volume of the language (4) will produce a relative touch, so that the voice signal of _ should have different output effects with different emotional parameters (4). The effect adjustment module 160 is responsible for adjusting the input and output of the voice signal corresponding to each character in the input parameter of the input module 12, and the effect adjustment module (10) in the voice signal is changed and changed in the voice signal. The amplitude of the voice signal is equal to the output of the voice signal, such as the production of the material, shooting, light, fade, off or change. Among them, the effect adjustment module (10) does not adjust the output effect of all the voice signals, and only adjusts the voice signal that needs to adjust the output effect. For example, the vibrato effect usually only appears in the length of the sound. The tone adjustment module 160 only adjusts the length of the played play beyond a certain voice signal to an output effect with vibrato. Further, the operation system and method of the present invention will be explained by way of an embodiment, and the flow chart of the simulation method for simulating vocal singing proposed by the present invention will be described with reference to "Fig. 3". In the present embodiment, the singing environment is provided to the user by the sing-along device 1 〇〇, the television (video playback device 6 〇〇), and the home audio (sound playback device 700) having the present invention. Before using the present invention, the storage module 110 on the accompaniment device 100 of the present invention first stores the voice files corresponding to all the Chinese characters (step 210). It is assumed that in the present embodiment, the Chinese character has a corresponding relationship with the 3 voice channel, and since the Chinese character has a basic syllable (word), at least 411 voice files are stored in the storage module 110. And different words and sounds correspond to at least one voice file. When the user turns on the function of the vocal guide in the singing environment, but the singer 100 of the present invention does not store the original sing of the song sung by the user, the input module 120 inputs the user in the singing environment. The lyrics of the song sung and the tune of the same song (step 221). Assuming that the selected song is "Zizhu Tune", the input module 120 to the storage module 110 reads the tune of the song "Zizhu Tune" and the lyrics of the song "Zizhu Tune" as shown in "Fig. 4". 3〇〇, to complete the song "Zizhu 15 1375214 tune" lyrics and the input of the tune. After that, the voice capturing module 130 reads the voice signal corresponding to each character in the lyrics 300 of the song "Zizhu Tun" in the storage module 11 (step 23G), because in this embodiment, each text system The word sound of the text corresponds to the voice slot, so the voice operation module 13() will separate the words in the lyrics 3〇〇, that is, the voice capture module 130 will follow the text. The word "one" is searched for the corresponding voice file' and the voice file corresponding to the text "one" is read. The voice signal corresponding to the word "one" is extracted from the voice file corresponding to the word "one". After the voice job corresponding to the text "_", continue to search for the voice file corresponding to the text "root", the text "purple", ..., the text "will" and the text "了" and take the voice signal from the voice file. Assuming that the plurality of voice files and texts stored in the storage module 110 are "root", the voice couching module 13G searches for a hidden Markov concept from a plurality of voice buildings corresponding to the "root" of the text. The most suitable voice file for voice 縦 module 15 () adjustment. The rest of the text can be analogized, so I won't go into details. After the θ extraction module 13 〇 to the storage module 摘 extracts the voice signal corresponding to each character in the lyrics (step 23G), the voice adjustment module 150 adjusts the voice operation module and captures the voice operation module. The pitch of the voice signal and the length of the voice signal are such that the voice of the voice is difficult to be the same as the tone of the voice towel, and the length of the voice signal is the same as the length of the tone (step 25〇) ^ For convenience of explanation, the tune "Wenzi" of the song "Zizhu Tune" will be described as "sol" in the tune as "sol" in the tune side as shown in "5th figure" and the play button is two shots. Therefore, the voice adjustment module 15G can adjust the pitch of the voice signal corresponding to the text "1" to "S01" using the time domain fundamental frequency synchronization superposition algorithm and adjust (extend or shorten) the playback length of the voice signal into two. Beat, but the voice adjustment module 150 is not limited to adjusting the voice signal using the time domain baseband synchronization plus algorithm. The manner in which the voice adjustment module 150 adjusts the voice signal corresponding to the rest of the characters in the lyrics 300 can be analogized, and therefore is not described. In the step of the voice capture module 130 extracting the voice signal (step 23〇) and the tone placement length of the voice module 15G voice field (step 250), the voice adjustment mode-group-(10) one After the voice signal corresponding to the text is taken out, the tone and the length of the voice ring can be adjusted immediately, and the voice capture module 130 can also correspond to all the characters in the voice capture module 130. After the voice signals are all taken out, the pitch and play length of all the extracted voice signals are adjusted separately. After the voice adjustment module 150 adjusts the pitch and play length of all the voice signals searched by the voice capture module 13 (step 250), the voice synthesis module 170 will sequentially follow the order of the characters in the lyrics. 1 to the corresponding sui audio signal, and after the completion of the synthesis of all the characters in the lyrics, the voice data is generated (step 270), that is, the voice synthesis module 17 will be based on the text "one" and the text "root", In the order of the words "purple", ..., the text "will" and the text "了", the voice signal θ corresponding to the text "root" is connected to the voice signal corresponding to the text "one", and the speech data is synthesized and corresponding to The voice signal of the word "purple" is connected to the synthesized voice data, the voice data of the f, ..., and so on, until the voice signal of the synthesized text "has" produces the most _ voice data 5 〇〇, the synthesized voice After the data is played, the playback length of each character in the lyrics 300 will be as shown in "Figure 6". After the speech synthesis module 170 synthesizes the speech signals corresponding to the characters to generate the speech data (step 270), the speech output module 18 outputs the synthesized speech data (step 280). It is assumed that in the embodiment, the voice output group 180 outputs voice data in a manner of playing voice data, and the singer device 1 executing the present invention directly plays the voice output module 18 and plays the stored message in the storage. The module hits 0-, and the melody of the lyrics is transmitted to the Jingyi audio. At the same time, the accompaniment device 100 executing the present invention also transmits the lyrics of the song to the television display, so that the melody of the song, the lyrics of the song, and the voice data are synchronized. The play allows the user to listen to the vocals that sing along with the tune of the tune in the singing environment, and to watch the lyrics of the humming. As described above, the present invention can further provide the effect adjustment module 160. After the effect adjustment module 16 is added, the user can input the lyrics 300 and the tune (step 221) through the input module 120. The emotion parameter is input through the input module 120 (step 223), and after the voice adjustment module 150 adjusts the pitch and play length of the voice signal (step 250), the effect adjustment module 160 adjusts the output effect of each voice signal according to the emotion parameter. (Step 260), for example, when the emotional parameter input by the user is "confidence", the effect 赃 module 16G can utter a longer voice signal according to the emotional parameter "confidence", for example, the first one of "straight seedlings" The last half of the voice signal corresponding to the word "Miao" is adjusted to vibrate, which is the pitch of the last H beat of the difficult voice signal, so that the sound is slightly off the original pitch; or the input is fine. When the emotional parameter is "angry", the effect adjustment template address 160 is based on the amplitude of the emotional parameter as the output of the "angry" difficult speech signal, that is, the sound of the adjusted speech signal. , But the present invention is referred to adjust the output effect is not limited to those in the above two. In the step that the voice adjustment module 150 adjusts the pitch and play length of the voice signal (step 250) and the effect adjustment module 16 adjusts the output effect of each voice signal (step 260), the effect adjustment module 16 can be used in the έ吾After the tone adjustment module 150 completes the adjustment of the length of the voice message, the voice signal output effect is adjusted immediately, and the voice adjustment module 15G can adjust the tone and the play length of all voice signals. After the completion, the voice signal of the New Fruit needs to be adjusted in the ride. After the effect adjustment module 160 adjusts the output effect of each voice signal (step 260), the voice synthesis module 17 合成 combines the voice signals adjusted by the voice adjustment module 150 and the effect adjustment module 160 into voice data (steps). 270), and the synthesized voice data is output by the voice output module 丨8〇 (step 280), so that the user can obtain the voice data of the singing voice by the voice. According to the above description, it can be seen that the difference between the present invention and the prior art is that it has a voice signal corresponding to each character in the lyrics of the song, and adjusts the pitch and length of the voice signal according to the tune of the song, and then according to The arrangement order of the characters in the lyrics is combined with the adjusted voice signals. By this technical means, the problem that the prior art cannot provide the vocal guide function for all the songs can be solved, thereby achieving the effect of simulating similar vocal singing.

Furthermore, the method for synthesizing vocal songs of the present invention can be implemented in a hardware software or a combination of hardware and software, or can be implemented in a centralized manner in a computer system or in a computer system with different components. Decentralized implementation.

Although the embodiments of the present invention are as described above, the contents are not limited to the scope of the present invention. Any person skilled in the art to which the present invention pertains has financial knowledge, and in the spirit and scope of the disclosure of the present invention, in the form of implementation and the fine red is the invention of the invention, the patent protection of the present invention The definition of the range = ___ is subject to change. Figure 1 is a schematic diagram of the environment for providing synthesized vocal songs according to the present invention. Fig. 2 is a diagram showing the flow of synthetic vocal music of the present invention. ° Fig. 4 is a lyrics of the embodiment of the present invention. Field 20 Figure 5 is a simplified representation of an embodiment of the invention. Figure 6 is a schematic diagram showing the length of sound playback of the voice data played by the embodiment of the present invention. [Main component symbol description] 100 vocal equipment 110 storage module 120 input module 130 voice capture module 150 voice adjustment module 160 effect adjustment module 'voice care' - one ---------- --- 180 Voice Output Module 300 Lyrics 400 Spread Spectrum 500 Voice Data 600 Video Playback Device 700 Sound Playback Device Step 210 Store the voice file of the corresponding text Step 221 Enter the song lyrics and song tune Step 223 Enter the emotion parameters Step 230纟The voice signal corresponding to each character in the corresponding lyrics of the lyrics is taken. Step 250 The pitch and play length of the voice signal are the same as the corresponding pitch and play length of the corresponding letter 1375214 in the tune. Step 260 is adjusted according to the emotional parameters. The voice signal output effect step 270 synthesizes the corresponding voice signals into voice data according to the order of the characters in the lyrics. Step 280 outputs the voice data 22

Claims (1)

X. Patent application scope: L A method for simulating vocal singing, which is applied to a accompaniment device, the accompaniment device outputs one lyric of a song and one tune of the song, and the method comprises the following steps: storing a plurality of corresponding at least a voice file of a character; inputting the lyrics and the tune; by the voice files corresponding to the texts in the lyrics; capturing respective voice signals corresponding to the respective characters; adjusting the pitch and playback of each voice signal The length is the same as the corresponding pitch and play length of the corresponding text in the tune; according to the order of the text in the lyric, the corresponding voice signals are synthesized as a voice data; and the voice data is output. 2. The method for simulating vocal singing as described in claim 1, wherein the step of extracting the voice signals further comprises searching for the voice files by a Hidden Markov Model (HMM). . 3. The method for simulating vocal singing as described in the scope of the patent application, in which the steps of extracting the voice signals from the voice files are respectively taken out from the voice files Part of this voice signal. 4. The method for simulating vocal singing according to the scope of the patent application, wherein the step of adjusting the pitch of each of the voice signals is synchronized with the time domain base frequency 23 1375214 (Time Domain Pitch Synchronous 〇veriap_add TD-PSOLA) The algorithm adjusts each of the voice signals. 5. The method of simulating vocal singing according to claim 1, wherein the method further comprises: providing an input of an emotional parameter after adjusting the pitch of each of the voice signals, and adjusting the emotion parameters according to the emotional parameter The steps of the output of the voice signal. 6. The method of simulating vocal singing according to item 5 of the patent application, wherein the step of adjusting the output effect of the voice signals comprises adjusting at least one of a pitch and a volume of the voice signal. 7. A system for simulating vocal singing, applied to a accompaniment device, which outputs a lyric of a song face and a tune of the song, the system comprising: a storage module for storing a plurality of a voice file, each of the voice files corresponding to at least one character; an input module for inputting the lyrics and the tune; a voice capture module for using the voice files corresponding to each text in the lyrics The voice adjustment module is configured to adjust the pitch and the play length of each voice signal to be the same as the corresponding pitch and play length of the corresponding text in the tune; a speech synthesis module for synthesizing the speech signals according to the order of the characters in the lyrics; and IS1 24 a speech transmission domain group '(4) outputting the speech data. 8. t Apply for the 7th reward of the simulated vocal singing system, "中.... J spider group uses the hidden model to search for the corresponding voice files of these words. 9. If you apply for the general (4) 7 items The system for simulating vocal singing, wherein the voice operation module is further configured to respectively extract the corresponding part of the text into the voice signals from the voice structures. • 1G. The system for simulating vocal singing, wherein the voice adjustment module adjusts the 5 sin-song signal by a time-domain fundamental frequency synchronous superposition algorithm. 11. The vocal vocal vocalization as described in the application of the syllabus _ 7 The system further includes an effect adjustment module for adjusting an output effect of the voice signals according to an emotional parameter, wherein the emotion parameter is input by the wheel entry module. The system for simulating vocal singing described in item u, wherein the output effect is at least one of vibrato, vibrato, fade in, fade out, delay or volume. 13· Simulated vocal singing as described in claim 11 System, The effect of the 5-week module is to adjust the voice according to the emotional parameter. The pitch and volume of the signal are at least one of them.